acase@tulane.edu

acase@tulane.edu

acase@tulane.edu

In this dissertation the results of two lateral load tests, performed at Chalmette, Louisiana for the Louisiana Power and Light Company, is presented. The two instrumented test piles were both forty feet long. However, the geometry of the two piles was different, resulting in piles having differing stiffness and resistability to lateral loads. One pile was fabricated without wings. The other, the rocket shaped pile, had wings attached to the upper twenty feet The soil at the test site generally consisted of fifteen feet of very soft organic clay and peat, overlying very soft to soft highly plastic clay. The water table was above the ground surface. The lateral load was applied to the foundation pile at fifty nine feet above ground line, simulating hurricane generated loads on tower structures. The loading procedure followed in the testing program included (1) short-term static loading, (2) cyclic loading, (3) unloading and reloading, and (4) load to failure The reaction of the test piles was predicted, utilizing an available finite difference computer program and procedures for constructing soil resistance and pile deflection (p-y) curves. In situ-p-y curves were developed from the field data and compared to those calculated from laboratory compression tests. Correlations between the in situ p-y curves and experimental curves were made and a new method for constructing p-y curves for very soft cohesive soil is presented. Other subjects addressed include (1) the relative insensitivity of highly plastic, very soft soils to cyclic loading, (2) the effect of unloading the test piles and subsequent reloading, (3) the effectiveness of the wings on the upper twenty foot length, and (4) design optimization / acase@tulane.edu

acase@tulane.edu

Mathematical models of the transport of oxygen and the simultaneous transport of carbon dioxide, glucose and lactic acid within the microcirculation of the human brain have been developed. These were solved for both steady-state and dynamic cases which included normal and pathological conditions The models are distributed parameter, interactive systems of non-linear partial differential equations. These were reduced to finite difference equations and solved using various numerical techniques. All three models employ the Price-Varga-Warren finite difference approximations for the capillary equations. The steady-state constant tissue metabolism model (SSCM) requires an analytical steady-state radial mass flux from the capillary. The dynamic non-linear tissue metabolism model (DNLM) uses Crank-Nicolson analogs in solving the non-linear, radial diffusion tissue equations. The third model is similar to the DNLM model but also includes axial diffusion in the tissue. Both the Extrapolated Liebmann and the Alternating Direction Implicit methods were used to solve the tissue equations of this model. Steady-state overall mass balances for normal conditions revealed errors of 0.1, 2.0 and 12.0 percent, respectively, for each of the three models The models incorporate in the transport description, the processes associated with the interaction of components through the Bohr and Haldane effects and the non-linear metabolism. Also studied was the oxygen mass transfer coefficient of the red blood cell and its effect on tissue oxygen delivery. Other cases investigated include normal conditions, arterial and venous hypoxia, reduced flow and hematocrit, hypocapnia, reduced glucose and with the dynamic models transient arterial upsets Results indicate that local concentration deficits of oxygen and/or glucose and excesses of lactic acid can exist with the system described by the models and therefore possibly could exist in the tissues of the human brain under similar circumstances. The implications are that the local energy production may be impaired directly or indirectly by the lack of oxygen and/or glucose and the existence of a local acidic environment. It is also proposed that local variations of components can be masked in experimental studies of regions of multiple capillaries, since the total chemical content gives no indication of local deficits and/or excesses / acase@tulane.edu

The objectives of this investigation were to measure the effect of the various factors that contribute to differential-pressure pipe sticking and to utilize this information in the development of a procedure to select casing setting depths The investigation differs from other investigations in that it (1) evaluates several areas in the Gulf of Mexico, not just one specific area; (2) evaluates both deviated and straight holes; (3) analyzes both wells that experience and did not experience sticking problems; and (4) analyzes the factors that may contribute to differential-pressure pipe sticking The results of this investigation normalized all factors that contribute to differential-pressure pipe sticking except differential pressure. It was concluded that the average differential pressure for those wells that experienced sticking problems and those that did not were not similar. It was also concluded that, of the wells that experienced sticking problems, more than 99 percent of them were stuck at pressures greater than 1,298 pounds per square inch The major application of this investigation is a procedure to select casing setting depths based on the ability to withstand a design-size kick and the potential of pipe sticking off-bottom / acase@tulane.edu

acase@tulane.edu

acase@tulane.edu

The purpose of this research project was to improve the accuracy of the predictions for the solubility of hydrocarbon solids in supercritical fluids. The role of size and energy asymmetries of solid-supercritical fluid systems on the attractive and repulsive pressure terms of various equations of state was determined. Modifications to improve the accuracy of the solubility predictions can be made based on the role of these pressure terms in the model The Peng-Robinson equation of state, with different random fluid mixing rules, was used to predict the solubility of a solid in a supercritical solvent. Discrepancies in accurately predicting the solubility appear to be due to the random fluid assumption rather than the form of the particular mixing rule. A modification of the van der Waals one-fluid mixing rule which includes both temperature and density dependence for the attractive parameter of the equation of state was developed based on a probabilistic approach. With the proposed modification, the Peng-Robinson equation of state can provide good solubility predictions Discrepancies in accurately predicting the solubility of solids in supercritical fluids with the more complex Boublik-Alder-Chen-Kreglewski equation of state are again due to the random fluid approximation in the mixing rule. The Perturbed-Hard-Chain equation of state, using rigorously derived mixing rules and eliminating the random fluid assumption (a nonrandomness concept), yields good results for solubility predictions. Importantly, the Perturbed-Hard-Chain approach offers the ability of extrapolative calculations, since the binary interaction coefficient can be obtained from previously compiled data. To give a more theoretical basis for the dependence of the attractive term of the equation of state on density, Local Composition Theory was used with the three equations of state, but its application yielded poor results for solubility predictions To determine the role of the repulsive force contribution to the equation of state, the combinatorial entropy of mixing for the system was calculated using the repulsive term of the three equations of state. Results suggest that the repulsive term of the Perturbed-Hard-Chain equation of state is the most correct form / acase@tulane.edu